As much as I think that’s correct a lot of the time, something like Bruno has value too. Implementing complicated auth for an annoying service once and reusing it across several pre-written requests, useful features like a GUI and history to see prior responses from an endpoint, being able to share the “collection” in the repo as examples/developer tools that’s maintained alongside the code, writing docs with each request to explain its usage, this stuff does add value that isn’t trivial to do with curl.

It’s usually pretty good about that, very apologetic (which is annoying), and usually does a good job taking it into account, although it sometimes needs reminders as that “context” gets lost in later messages.

I’ll give some examples. In that same networking session, it disabled some security feature, to test if it was related. It never remembered to turn that back on until I specifically asked it to re-enable “that thing you disabled earlier”. To which it responds something like “Of course, you’re right! Let’s do that now!”. So, helpful tone, “knew” how to do it, but needed human oversight or it would have “forgotten” entirely.

Same tone when I’d tell it something like “stop starting all your commands with SSH, I’m in an SSH session already.” Something like “of course, that makes sense, I’ll stop appending SSH immediately”. And that sticks, I assume because it sees itself not using SSH in its own messages, thereby “reminding” itself.

Its usual tone is always overly apologetic, flattering, etc. For example, if I tell it bluntly I’m not giving my security credentials to an LLM, it’ll always say something along the lines of “great idea! That’s a good security practice”, despite directly suggesting the opposite moments prior. Of course, as we’ve seen with lots of examples, it will take that tone even if actually can’t do what you’re asking, such as in the examples of asking ChatGPT to give you a picture of a “glass of wine filled to the very top”, so it’s “tone” isn’t really something you can rely on as to whether or not it can actually correct the mistake. It’s always willing to take another attempt, but I haven’t found it always capable of solving the issue, even with direction.

Man, AI agents are remarkably bad at “self-awareness” like this, I’ve used it to configure some networking on a Raspberry Pi, and found myself reminding it frequently, “hey buddy, maybe don’t lock us out of connecting to this thing over the network, I really don’t want to have to wipe the thing because it’s running a headless OS”.

It’s a perfect example of the kind of thing that “walk or drive to wash your car?” captures. I need you to realize some non-explicit context and make some basic logical inferences before you can be even remotely trusted to do anything important without very close expert supervision, a degree of supervision that almost makes it totally worthless for that kind of task because the expert could just do it instead.

Solid advice, I’ve been using the “restore checkpoint” feature to go back to before it started trying an approach without totally tossing my context, and it’s certainly easier than arguing with it or ignoring its reminders to try X.

I’ve definitely seen it be stubborn like that in my tinkering with it, just absolutely locked on to a specific approach like a dog with a bone, even after I’ve already started nudging it to move on and try something else. I assume that’s a result of “recency bias” in its memory, missing the forest for the trees, because I don’t need that solution to work, I need a solution to my original problem, preferably the most elegant and least hacky solution.

Certainly one of the things that indicates to me that LLMs will be best used by someone who knows what they’re doing for the foreseeable future. Shame it also creates so much deskilling and discourages learning those skills in the first place. Absolutely something that worries me for our future.

Diagram in the link preview (and also in the article) is a little confusing, because “invaluable” subverts the usual meaning of the “in” prefix, with “invaluable” meaning:

extremely useful; indispensable

Which makes the diagram suggest that deep focus gets good results, but unfocused and distracted work gets the best results.

No kidding, it’s intense. I like the channel well enough, and find they’ve handled their various scandals to my satisfaction, at least. And yes, it’s 100% more entertainment than deep technical content I’m gonna learn from, but whatever, guilty pleasures aren’t some horrible sin, and it’s a better use of my time than reality TV. Being such a large channel does also get them access to some really worthwhile stuff now and again, like major factory tours, or Linus Torvalds, and that is stuff that’s really worth my time.

All of that said, I assume there’s lots of people here like myself, but it just ain’t worth jumping out of the woodwork for the guy. I like him, but not nearly enough to cast myself against this much vitriol for his sake.

This is the coolest video they’ve done in ages, and Linus Torvalds seems like such an excellent, genuine guy. He gives really thoughtful answers to even some of the silly joke questions, and it’s great. I super appreciate his perspective, and he interviews excellently.

Haha, I found myself thinking the same thing, and then caught myself, realizing all the other LLMs on this page had lowered the bar immensely for what I’m considering impressive.

Interesting stuff. Yeah, can confirm I’ve not had any experiences like that in my 6-ish months with it, despite screwing around with nearly everything under the sun: emulators, modding games, hosting services, third party launchers, etc, but I guess it shouldn’t be a huge surprise that it hasn’t always been that rock solid.

My only real issue so far has been that Steam isn’t quite wayland-ready, and I’m insistent on tinkering with HDR gaming and therefore run into issues with Steam Input or Steam Overlay.

Fascinating, what were you up to manage that? As far as I’m aware, that should basically only be possible with rpm-ostree commands, which are pretty strongly recommended against in the documentation, and I’ve yet to find something I really need any for.

Mhm, fair point. Although… I would say the steam deck’s popularity and proof of viability as a gaming device is doing an immense amount of work on its own. I built a gaming PC ~2 years ago, and even as a long time developer and someone comfortable with a UNIX terminal I opted to get a copy of Windows for gaming, and had to awkwardly get to grips with it and find tools to get it playing the way I wanted.

It’s only ~1 month ago that the prevalence and maturity of the steam deck (combined with Windows recall re-emerging🤮) finally had me at ease enough to give Bazzite a shot, and since jumping myself and expressing how happy I am with it, 2 of my long term “on the fence” friends have asked me questions and are starting to try Linux themselves.

Larger Linux market share, regardless of how it gets there, gives broad confidence in Linux, and also pushes developers and Steam itself to maintain Linux support and tools like Proton, which reinforces the cycle, even if it doesn’t help us “kill Windows” for as long as users don’t understand how to install it.

Might be a good use case for Anubis, in addition to the URLParam passwords mentioned elsewhere. Enough protection to prevent trivial brute force scraping, while also being basically invisible to users.